SmartForceps Data Framework for Glioma Surgery
Supplementary Document
Abdulrahman Albakr1
Amir Baghdadi2
Rahul Singh3
Sanju Lama4
Garnette R. Sutherland5
September 30, 2021
Force data from Expert and Novice surgeons; data extracted from 17 procedures involving Glioma tumor (multiple tasks per procedure) and segmented for each microsurgical task.
This document contains the data management and analytic framework for SmartForceps, a sensorized surgical bipolar forceps capable of quantifying the forces of tool-tissue interaction in microsurgery. We have shown that high force error is associated with bleeding, low force error with the need to repeat the task and force variability with both. We have further established that novice and intermediate level surgeons exert more force errors compared to experienced. The technology introduces a data-driven surgical paradigm whereby forces of tool-tissue interaction are used as an objective assessment metric for surgical competency. Together with the error warning system, the reduction of surgical error and improved patient safety is promised.
The present document incorporates 17 cases of neurosurgery performed at Foothills Medical Centre, Calgary. The data were extracted and segmented for each task of microsurgery performed by an Expert and multiple Novice surgeons.
The multimedia supplementary files for the software and device structure is available at https://github.com/smartforceps/supplementary/.
The reader can show any code chunk by clicking on the code button. We chose to make the default for the code hidden since we: (a) wanted to improve the readability of this document; and (b) assumed that the readers will not be interested in reading every code chunk.
The snippet below documents the list of R libraries that were used in this research. For convenience, we used the pacman package since it allows for installing/loading the needed libraries in one step.
rm(list = ls()) # clear global environment
graphics.off() # close all graphics
library(pacman) # needs to be installed first
p_load(
pastecs,
reticulate,
readr,
htmltools,
vembedr,
devtools,
usethis,
slam,
reshape2,
data.table,
gridExtra,
extrafont,
ISLR,
jpeg,
coda,
abind,
chron,
fmsb,
gdata,
stringr,
lubridate,
ggplot2,
ggpubr,
gghighlight,
ggridges,
plotly,
tidyverse,
fBasics,
signal,
GeneCycle,
Rwave,
seewave,
spectral,
tsfeatures,
peakPick,
pracma,
foreach,
glue,
dplyr,
rstatix,
emmeans,
EnvStats,
kableExtra
)Data Recording
Data is generated through strain gauge sensors placed along the prongs of a bipolar forceps. The technology is designed and manufactured within OrbSurgical, ltd. in partnership with Bissinger, Germany. The calibrated data is analyzed and recorded in realtime through our software platform which is overviewed in the snippet below. Please navigate through different items in the tabs for various pages in the software.
SmartForceps Software
Home Page
Adding New Case
Download Forceps and Hospital Data Page
Data Monitoring Page
Data Upload Page
Software Demo
Below is a short demo of our software showing the force profiles in realtime with a high force error warning system (with 0.8 N threshold) and a voice recognition module to identify the keywords for different surgical tasks.
Orbsurgical SmartForceps Web Application
SmartForceps comes with a web application to manage the data, device characteristics (e.g. calibration factors), and medical center and surgeon information. This platform is designed in a way to manage the data to and from the cloud through integration into the software.
Data Processing
Data was extracted and analyzed using R programming platform. In the snippet below, we import the “.txt” files obtained from our Azure data cloud and the “.xlsx” file containing each task information (e.g. time stamp, task name, surgeon name, and remarks of potential intraoperative force errors) extracted from the surgical team voice recorded during each case. The imported data are structured as R dataframes.
We have extracted various features from the segmented task force data in each prong. A combination of average, minimum, or maximum value of features for each prongs were included in the future analysis. These time-series features included:
Force Duration: duration of force application in one task segment.
Force Average: average of force values in one task segment.
Force Max: maximum of force values in one task segment.
Force Min: minimum of force values in one task segment.
Force Range: range of force values in one task segment.
Force Median: median of force values in one task segment.
Force SD: standard deviation of force values in one task segment.
Force CV: coefficient of variation of force values in one task segment.
Force Peak Value: peak force value in one task segment.
Force Peak Counts: number of force peaks in one task segment.
Force Signal Flat Spots: maximum run length for each section of force time-series when divided into ten equal-sized intervals.
Force Signal Trend: force time-series trend in one task segment.
Force Signal Fluctuations: force time-series fluctuation index in one task segment.
Force Signal Spikiness: force time series spikiness index (variance of the leave-one-out variances of the remainder component) in one task segment.
Force Signal Stability: force time-series stability index (variance of the means) in one task segment.
Force Signal Entropy: force time-series forecastability in one task segment (low values indicate a high signal-to-noise ratio).
Among these feature list, the best subset will be selected for the subsequent analysis based on statistical tests to monitor their representation power in different surgeon skill and task categories. The aim was to have the best explain of the patterns and behaviors for force profiles over the timespan of each data segment. To find accurate force peaks within each task segment, the signals were smoothed by passing through a digital 4st order Butterworth low-pass filter with a cutoff frequency of 0.1 Hz. Further, the outlier segmented data were identified based on 1st and 99th percentiles of either maximum force, minimum force or task completion time from all trials of the expert surgeon as <1% error was assumed to occur by experienced surgeons. The force segments for which the maximum force peak, minimum force valley, or task completion time exceeded the upper threshold (99th percentile) or fell short of the lower threshold (1st percentile) were labeled as outliers and removed (~11% were removed). The clean engineered features serve as a baseline for hand-crafted feature-based surgical skill classification.
The surgical tasks were classified as 5 main categories:
Retracting
Manipulation
Dissecting
Pulling
Coagulation
# defining functions
absmax <- function(x) { x[which.max( abs(x) )]}
# reading data
data_info <- read.xls ("~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Glioma Data Info - Subtask.xlsx",
sheet = 1, header = TRUE)
data_dir <- "~/Desktop/Canada/neuroArm/SmartForceps/SmartForceps Data"
cases <- c(5, 6, 7, 10, 15, 16, 22, 37, 42, 48, 53, 54, 56, 66, 77, 91) # Glioma Cases
# "SF-3": 2:23 (3.46, 3.90)
# "SF1-2020" 24:45, 47:51 (3.45, 3.45)
# "SF3-2020-S" 46 (3.90, 3.90)
forceps_names <- c(rep("SF-3", 22), rep("SF1-2020", 67), rep("SF3-2020-S", 1))
forceps_cases <- vector(mode="list", length=90)
names(forceps_cases) <- c(as.character(c(2:23)), as.character(c(24:45, 47:91)), as.character(46))
for (i in 1:length(forceps_cases)) {
forceps_cases[[i]] <- forceps_names[i]
}
calibrtion_factors <- list("SF-3" = c(3.46, 3.90),
"SF1-2020" = c(3.45, 3.45),
"SF3-2020-S" = c(4.43, 4.43))
# preparing each case data
for (i in cases) {
file_dir <- paste(c(data_dir, paste(c("Case", i), collapse = " "), "log data"), collapse ="/")
temp <- list.files(path = file_dir, pattern = "*.txt")
my_data <- data.frame()
for (j in 1:length(temp)) {
read_txt <- read.delim(paste(c(file_dir, temp[j]), collapse = "/"))
my_data <- rbind(my_data,
cbind(data.frame(DataSection = rep(j, dim(read_txt)[1])),
read_txt))
# apply calibration factors
my_data$LeftCalibratedForceValue <-
my_data$LeftRawVoltageValue*calibrtion_factors[[forceps_cases[[as.character(i)]]]]
my_data$RightCalibratedForceValue <-
my_data$RightRawVoltageValue*calibrtion_factors[[forceps_cases[[as.character(i)]]]]
assign(paste0("case_", i, "_forcedata"), my_data)
}
}
# concatenating case data
force_seg_data <- data.frame()
force_seg_info <- data.frame()
task_count <- c(0,0,0,0,0,0,0,0,0,0,0,0,0)
outlier_count <- 0
seg_num_res <- 0
acqFreq <- 20
for (i in 1:length(cases)) {
#print("Case:")
#print(cases[i])
case_i_info <- data_info[data_info[, "Case"] == cases[i],]
case_i_data <- get(paste(c("case", cases[i], "forcedata"), collapse = "_"))
case_i_secpowerupstartidxs <- c(1, which(diff(case_i_data$DataSection) != 0) + 1)
case_i_secpoweruptimes <- case_i_data$MillisecondsSincePowerUp[case_i_secpowerupstartidxs]
for (j in 1:dim(case_i_info)[1]) {
data_section <-
as.numeric(strsplit(toString(case_i_info$Remarks[j]), split = " ")[[1]][2])
if(is.na(strsplit(toString(case_i_info$Remarks[j]), split = " ")[[1]][5])){
surgeon_experience <- "Expert"
} else {
surgeon_experience <- "Novice"
}
if(size(strsplit(toString(case_i_info$Remarks[j]), split = " ")[[1]])[2] > 2){
surgeon_name <- tail(strsplit(toString(case_i_info$Remarks[j]), split = " ")[[1]], n=1)
} else {
surgeon_name <- "Dr. Sutherland"
}
task_type <-
strsplit(str_sub(case_i_info$Task[j]), split = " ")[[1]][1]
if (task_type == "Coagulation"){
task_type <- str_sub(case_i_info$Task[j])
}
t_start <-
as.numeric(seconds(hms(case_i_info$TimeStart[j]))) * 1000 + case_i_secpoweruptimes[data_section]
t_end <-
as.numeric(seconds(hms(case_i_info$TimeEnd[j]))) * 1000 + case_i_secpoweruptimes[data_section]
idx_start <-
as.numeric(rownames(case_i_data[case_i_data[, "MillisecondsSincePowerUp"] == t_start, ]))
idx_end <-
as.numeric(rownames(case_i_data[case_i_data[, "MillisecondsSincePowerUp"] == t_end, ]))
# LeftCalibratedForce <- case_i_data[idx_start:idx_end, "LeftCalibratedForceValue"]
# RightCalibratedForce <- case_i_data[idx_start:idx_end, "RightCalibratedForceValue"]
TimeS <- seq(0, length(idx_end:idx_start)*0.05-0.05, by=0.05)
# handle each task accordingly
if (strsplit(str_sub(case_i_info$Task[j]), split = " ")[[1]][1] == "Coagulation"){
# only positive values are expected
LeftCalibratedForce <- case_i_data[idx_start:idx_end, "LeftCalibratedForceValue"] -
min(case_i_data[idx_start:idx_end, "LeftCalibratedForceValue"])
RightCalibratedForce <- case_i_data[idx_start:idx_end, "RightCalibratedForceValue"] -
min(case_i_data[idx_start:idx_end, "RightCalibratedForceValue"])
} else if (strsplit(str_sub(case_i_info$Task[j]), split = " ")[[1]][1] == "Dissecting"){
# only negative values are expected (returned to positive)
LeftCalibratedForce <- abs(case_i_data[idx_start:idx_end, "LeftCalibratedForceValue"] -
max(case_i_data[idx_start:idx_end, "LeftCalibratedForceValue"]))
RightCalibratedForce <- abs(case_i_data[idx_start:idx_end, "RightCalibratedForceValue"] -
max(case_i_data[idx_start:idx_end, "RightCalibratedForceValue"]))
} else {
LeftCalibratedForce <- case_i_data[idx_start:idx_end, "LeftCalibratedForceValue"]
RightCalibratedForce <- case_i_data[idx_start:idx_end, "RightCalibratedForceValue"]
}
if (task_type == "Coagulation"){
task_count[1] = task_count[1]+1
seg_num_task = task_count[1]
} else if (task_type == "Coagulation (Vessel)"){
task_count[2] = task_count[2]+1
seg_num_task = task_count[2]
} else if (task_type == "Coagulation (Galea)"){
task_count[3] = task_count[3]+1
seg_num_task = task_count[3]
} else if (task_type == "Coagulation (Dura)"){
task_count[4] = task_count[4]+1
seg_num_task = task_count[4]
} else if (task_type == "Coagulation (Muscle)"){
task_count[5] = task_count[5]+1
seg_num_task = task_count[5]
} else if (task_type == "Coagulation (Pia / Arachnoid)"){
task_count[6] = task_count[6]+1
seg_num_task = task_count[6]
} else if (task_type == "Coagulation (Tumor)"){
task_count[7] = task_count[7]+1
seg_num_task = task_count[7]
} else if (task_type == "Coagulation (Brain)"){
task_count[8] = task_count[8]+1
seg_num_task = task_count[8]
} else if (task_type == "Coagulation (Gliotic / Scar Tissue)"){
task_count[9] = task_count[9]+1
seg_num_task = task_count[9]
} else if (task_type == "Pulling"){
task_count[10] = task_count[10]+1
seg_num_task = task_count[10]
} else if (task_type == "Manipulation"){
task_count[11] = task_count[11]+1
seg_num_task = task_count[11]
} else if (task_type == "Dissecting"){
task_count[12] = task_count[12]+1
seg_num_task = task_count[12]
} else if (task_type == "Retracting"){
task_count[13] = task_count[13]+1
seg_num_task = task_count[13]
}
# segment data
force_seg_data <- rbind(
force_seg_data,
data.frame(
"CaseNum" = rep(cases[i], length(idx_end:idx_start)),
"SegmentNum" = rep(j, length(idx_end:idx_start)),
"SegmentNumTask" = rep(seg_num_task, length(idx_end:idx_start)),
"SegmentNumOverall" = rep(seg_num_res+j, length(idx_end:idx_start)),
"Time" = TimeS,
"LeftForce" = LeftCalibratedForce,
"RightForce" = RightCalibratedForce,
"TaskType" = rep(task_type, length(idx_end:idx_start))
))
desc_stats_left <- stat.desc(LeftCalibratedForce, norm=TRUE)
desc_stats_right <- stat.desc(RightCalibratedForce, norm=TRUE)
forceDetrendLeft <- as.numeric(lm(LeftCalibratedForce ~TimeS)$residuals)
forceDetrendRight <- as.numeric(lm(RightCalibratedForce ~TimeS)$residuals)
# the first 10 harmonics of the signal
forceFreqLeft <- order(-Mod(fft(forceDetrendLeft))[1:(acqFreq/2)])-1
forceFreqRight <- order(-Mod(fft(forceDetrendRight))[1:(acqFreq/2)])-1
# converting signal to time series
tsForceLeft <- ts(LeftCalibratedForce, frequency = forceFreqLeft[1])
tsForceRight <- ts(RightCalibratedForce, frequency = forceFreqRight[1])
# extracting time series features
tsFeaturesLeft <- tsfeatures(LeftCalibratedForce)
tsFeaturesRight <- tsfeatures(RightCalibratedForce)
# filter force profiles to find peaks
bf <- butter(4, 0.1, type="low")
LForceFilt <- signal::filter(bf, LeftCalibratedForce)
# plot(LeftCalibratedForce, type = "l")
# lines(LForceFilt, col = "red")
peakhits <- peakpick(matrix(LForceFilt, ncol=1), neighlim=5, peak.min.sd=5, peak.npos=20)
# peaks from mean
forcePeaksLeft <- LForceFilt[peakhits] - mean(LForceFilt)
if (length(forcePeaksLeft) == 0) {
forcePeaksLeft <- NA
}
# forcePeaksLeft <- findpeaks(as.numeric(LForceFilt), minpeakdistance = 5)
# plot(LForceFilt)
# points((1:length(LForceFilt))[peakhits], LForceFilt[peakhits], col="red")
RForceFilt <- signal::filter(bf, RightCalibratedForce)
#plot(RightCalibratedForce, type = "l")
#lines(RForceFilt, col = "red")
peakhits <- peakpick(matrix(RForceFilt, ncol=1), neighlim=5, peak.min.sd=5, peak.npos=20)
# peaks from mean
forcePeaksRight <- RForceFilt[peakhits] - mean(RForceFilt)
if (length(forcePeaksRight) == 0) {
forcePeaksRight <- NA
}
# forcePeaksRight <- findpeaks(as.numeric(RForceFilt), minpeakdistance = 5)
# plot(RForceFilt)
# points((1:length(RForceFilt))[peakhits], RForceFilt[peakhits], col="red")
# handle no peak instances
if (is.na(forcePeaksLeft) & is.na(forcePeaksRight)) {
forcePeaks <- NA
} else {
forcePeaks <- max(max(forcePeaksLeft, na.rm = TRUE),
max(forcePeaksRight, na.rm = TRUE),
na.rm = TRUE)
}
# plot.frequency.spectrum <- function(X.k, xlimits=c(0,length(X.k))) {
# plot.data <- cbind(0:(length(X.k)-1), Mod(X.k))
#
# # TODO: why this scaling is necessary?
# plot.data[2:length(X.k),2] <- 2*plot.data[2:length(X.k),2]
#
# plot(plot.data, t="h", lwd=2, main="",
# xlab="Frequency (Hz)", ylab="Strength",
# xlim=xlimits, ylim=c(0,max(Mod(plot.data[,2]))))
# }
# segment info
force_seg_info <- rbind(
force_seg_info,
data.frame(
"CaseNum" = cases[i],
"SegmentNum" = j,
"SegmentNumTask" = seg_num_task,
"SegmentNumOverall" = seg_num_res+j,
"TaskType" = task_type,
"User" = surgeon_experience,
"SurgeonName" = surgeon_name,
"Duration.Force" = round(length(idx_end:idx_start)*0.05-0.05, digits=4),
"Mean.Force" = round(mean(desc_stats_left['mean'],
desc_stats_right['mean']), digits=4),
# "Mean.Force" = round(mean(desc_stats_left['range'],
# desc_stats_right['range'])/2, digits=4),
"Max.Force" = round(max(desc_stats_left['max'],
desc_stats_right['max']), digits=4),
"Min.Force" = round(min(desc_stats_left['min'],
desc_stats_right['min']), digits=4),
"Range.Force" = round(max(desc_stats_left['range'],
desc_stats_right['range']), digits=4),
"Median" = round(median(desc_stats_left['median'],
desc_stats_right['median']), digits=4),
"SD" = round(max(desc_stats_left['std.dev'],
desc_stats_right['std.dev']), digits=4),
"Coef.Variance" = round(max(desc_stats_left['coef.var'],
desc_stats_right['coef.var']), digits=4),
"Force.Peaks" = round(forcePeaks, digits=4),
"Peaks.Count" = round(max(length(forcePeaksLeft),
length(forcePeaksRight)), digits=4),
"Flat.Spots" = round(min(flat_spots(LeftCalibratedForce),
flat_spots(RightCalibratedForce)), digits=4),
"Trend" = round(mean(tsFeaturesLeft[['trend']],
tsFeaturesRight[['trend']]), digits=4),
"Fluctuation" = round(absmax(c(fluctanal_prop_r1(LeftCalibratedForce),
fluctanal_prop_r1(RightCalibratedForce))), digits=4),
"Spike" = round(absmax(c(tsFeaturesLeft[['spike']],
tsFeaturesRight[['spike']])), digits=4),
"Stability" = round(mean(stability(LeftCalibratedForce),
stability(RightCalibratedForce)), digits=4),
"Entropy" = round(max(tsFeaturesLeft[['entropy']],
tsFeaturesRight[['entropy']]), digits=4)
))
}
seg_num_res <- seg_num_res+dim(case_i_info)[1]
}
row.names(force_seg_info) <- 1:dim(force_seg_info)[1]
# filtering the outliers: cutoffValues = [time, max value, min value, task id]
tasks = c("Coagulation", "Coagulation (Vessel)", "Coagulation (Galea)",
"Coagulation (Dura)", "Coagulation (Muscle)", "Coagulation (Pia / Arachnoid)",
"Coagulation (Tumor)", "Coagulation (Brain)", "Coagulation (Gliotic / Scar Tissue)",
"Pulling", "Manipulation", "Dissecting", "Retracting")
features = c("Duration.Force",
"Min.Force",
"Max.Force")
thresholds_1 <- vector()
thresholds_99 <- vector()
for (i in 1:length(features)) {
thresholds_tasks_1 <- vector()
thresholds_tasks_99 <- vector()
for (k in 1:length(tasks)) {
data <- force_seg_info %>%
filter(TaskType == tasks[k],
User == "Expert")
data <- data[features[i]]
threshold <- quantile(data[[1]], c(.01, .99))
thresholds_tasks_1 <- rbind(thresholds_tasks_1, threshold['1%'][[1]])
thresholds_tasks_99 <- rbind(thresholds_tasks_99, threshold['99%'][[1]])
}
thresholds_1 <- cbind(thresholds_1, thresholds_tasks_1)
thresholds_99 <- cbind(thresholds_99, thresholds_tasks_99)
}
df_thresholds_1 <- data.frame(thresholds_1)
colnames(df_thresholds_1) <- features
rownames(df_thresholds_1) <- tasks
df_thresholds_99 <- data.frame(thresholds_99)
colnames(df_thresholds_99) <- features
rownames(df_thresholds_99) <- tasks
outlier_count <- 0
force_seg_info_outlier_removed <- data.frame()
for (i in 1:nrow(force_seg_info)) {
segDur <- force_seg_info[i, 'Duration.Force']
segMin <- force_seg_info[i, 'Min.Force']
segMax <- force_seg_info[i, 'Max.Force']
if(sum(df_thresholds_99[force_seg_info[i, 'TaskType'], ] < c(segDur, segMin, segMax)) >= 1){
# print("outlier detected")
outlier_count <- outlier_count + 1
}
else if (sum(c(segDur, segMin, segMax) < df_thresholds_1[force_seg_info[i, 'TaskType'], ]) >= 1){
# print("outlier detected")
outlier_count <- outlier_count + 1
}
else {
force_seg_info_outlier_removed <- rbind(force_seg_info_outlier_removed, force_seg_info[i, ])
}
}
save(
case_5_forcedata,
case_6_forcedata,
case_7_forcedata,
case_10_forcedata,
case_15_forcedata,
case_16_forcedata,
case_22_forcedata,
case_37_forcedata,
case_42_forcedata,
case_48_forcedata,
case_53_forcedata,
case_54_forcedata,
case_56_forcedata,
case_66_forcedata,
case_77_forcedata,
case_91_forcedata,
file = "~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Subtasks Included/SmartForcepsDataRead.RData")
save(
force_seg_data,
file = "~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Subtasks Included/SmartForcepsDataProcessed.RData"
)
save(
force_seg_info,
file = "~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Subtasks Included/SmartForcepsDataFeature.RData"
)
save(
force_seg_info_outlier_removed,
file = "~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Subtasks Included/SmartForcepsDataFeatureClean.RData"
)
df_feature <- melt(data = force_seg_info_outlier_removed,
id.vars = c("CaseNum",
"SegmentNum",
"SegmentNumTask",
"SegmentNumOverall",
"TaskType",
"User",
"SurgeonName"),
measure.vars = c("Duration.Force",
"Mean.Force",
"Max.Force",
"Min.Force",
"Range.Force",
"Median",
"SD",
"Coef.Variance",
"Force.Peaks",
"Peaks.Count",
"Flat.Spots",
"Trend",
"Fluctuation",
"Spike",
"Stability",
"Entropy"),
variable.name = "FeatureName",
value.name = "Value")
df_feature <- data.frame(lapply(df_feature, as.character), stringsAsFactors=FALSE)
df_processed <- melt(data = force_seg_data,
id.vars = c("CaseNum",
"SegmentNum",
"SegmentNumTask",
"SegmentNumOverall",
"TaskType",
"Time"),
measure.vars = c("LeftForce", "RightForce"),
variable.name = "ProngName",
value.name = "Value")
# returnOptions <- function(feature_data){
# features <- feature_data$FeatureName %>% unique(.)
# foreach(i=features) %do% list('label'=glue('feature: {i}'), 'value'=i)
# }
#
# default_featureid <- "DurationForce"
# default_feature_input <- df_feature[df_feature$FeatureName==default_featureid, ]
# default_options <- returnOptions(df_feature)
write.csv(df_processed,
'~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Subtasks Included/SmartForcepsDataProcessed.csv')
write.csv(df_feature,
'~/Desktop/Canada/neuroArm/SmartForceps/Glioma/codes/Subtasks Included/SmartForcepsDataFeature.csv')Force Profiles
Below are the interactive figures of the force profiles for all 17 cases categorized in 13 common surgical tasks categories. The graph can highlight the differences in completion time and range of forces across the 13 surgical tasks.
SmartForceps Left and Right prong data output:
Average of Right and Left Prong Force
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation",]
fig_c <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_c <- add_trace(fig_c,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_c <- fig_c %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_c <- fig_c %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Vessel)",]
fig_cv <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cv <- add_trace(fig_cv,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cv <- fig_cv %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cv <- fig_cv %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Vessel) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Galea)",]
fig_cg <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cg <- add_trace(fig_cg,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cg <- fig_cg %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cg <- fig_cg %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Galea) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Muscle)",]
fig_cm <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cm <- add_trace(fig_cm,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cm <- fig_cm %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cm <- fig_cm %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Muscle) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Dura)",]
fig_cd <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cd <- add_trace(fig_cd,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cd <- fig_cd %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cd <- fig_cd %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Dura) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Pia / Arachnoid)",]
fig_cpa <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cpa <- add_trace(fig_cpa,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cpa <- fig_cpa %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cpa <- fig_cpa %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Pia / Arachnoid) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Tumor)",]
fig_ct <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_ct <- add_trace(fig_ct,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_ct <- fig_ct %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_ct <- fig_ct %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Tumor) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Brain)",]
fig_cb <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cb <- add_trace(fig_cb,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cb <- fig_cb %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cb <- fig_cb %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Brain) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Coagulation (Gliotic / Scar Tissue)",]
fig_cgst <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_cgst <- add_trace(fig_cgst,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_cgst <- fig_cgst %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task E \n Froce (N) \n '))
fig_cgst <- fig_cgst %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Coagulation \n (Gliotic / Scar Tissue) \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Pulling",]
fig_p <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_p <- add_trace(fig_p,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_p <- fig_p %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task D \n Froce (N) \n '))
fig_p <- fig_p %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Pulling \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Manipulation",]
fig_m <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_m <- add_trace(fig_m,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_m <- fig_m %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task B \n Froce (N) \n '))
fig_m <- fig_m %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Manipulation \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Dissecting",]
fig_d <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_d <- add_trace(fig_d,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_d <- fig_d %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task C \n Froce (N) \n '))
fig_d <- fig_d %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Dissecting \n Froce (N) \n '))
data_select_task <- force_seg_data[force_seg_data$TaskType == "Retracting",]
fig_r <- plot_ly()
for (SegNum in 1:max(data_select_task$SegmentNumTask)) {
data_select_seg <- data_select_task[data_select_task$SegmentNumTask == SegNum, ]
fig_r <- add_trace(fig_r,
x = data_select_seg$Time,
y = rowMeans(data_select_seg[,c('LeftForce', 'RightForce')], na.rm=TRUE),
type = 'scatter', mode = 'lines',
fill = 'tozeroy', fillcolor = data_select_seg$SegmentNumTask,
showlegend = F,
line = list(width = 0.5))
}
# fig_r <- fig_r %>% layout(xaxis = list(title = 'Time (sec)'),
# yaxis = list(title = 'Task A \n Froce (N) \n'))
fig_r <- fig_r %>% layout(xaxis = list(title = 'Time (sec)'),
yaxis = list(range = c(-3, 4), title = 'Retracting \n Froce (N) \n '))
figs <- list(fig_r, fig_m, fig_d, fig_p, fig_c, fig_cv, fig_cg, fig_cd, fig_cpa, fig_ct, fig_cb, fig_cgst)
figs <- subplot(fig_r, fig_m, fig_d, fig_p, fig_c, fig_cv, fig_cg, fig_cd, fig_cpa, fig_ct, fig_cb, fig_cgst,
nrows = length(figs),
shareX = TRUE, titleX = TRUE, titleY = TRUE)
figs <- figs %>% layout(title = "Overlaid Force Signals Over Time <br> Based on Common Surgical Task Categories (Average of Right and Left Prongs)")
figs